INDUSTRY NEWS

THE MICRO INTERVIEW

Intel's Mike Splinter to suppliers: Improve your technology, lower your costs, get more agile

If you're talking about the brain trust responsible for copy exactly, virtual factory, and other concepts crucial to Intel's famed manufacturing prowess, one of the big kahunas is Michael R. "Mike" Splinter. Now a senior vp and general manager reporting directly to the executive triumvirate of Gordon Moore, Craig Barrett, and Andy Grove, he and Sunlin Chou run Intel's worldwide technology and manufacturing group. As the accompanying news story notes, the chipmaking giant plans to add significant manufacturing capacity during the next couple of years, including its first volume 300-mm plant, Fab 22, in Chandler, AZ. With shrinking technology cycles forcing semiconductor companies and their suppliers to be ready with fast-ramping tools and processes, Splinter's message to the equipment and materials community is clear: Be ready to quickly get the technology to us when we need you to, with cost-effective, more-productive products that help us achieve our manufacturing goals on time-- or else.

A graduate of the University of Wisconsin with BS and MS degrees in electrical engineering, Splinter worked for Rockwell before joining Intel in 1984 as manager of Fab 1 in Santa Clara, CA. He has moved steadily up through the ranks, assuming the assistant general manager position in the company's technology manufacturing group in October 1996 and getting the promotion to senior vp in January 1999. MICRO's Tom Cheyney caught up with Splinter in Pebble Beach after the Intel exec delivered the industry keynote address during SEMI's Industry Strategy Symposium in early January.

MICRO: How does Intel organize its yield teams or focus groups within the fabs and on the corporate level?

Splinter: Intel attacks yield from the virtual factory standpoint. Each factory certainly has a yield team that is working on parts of an overall problem and trying to reduce those defects. As they solve problems, those solutions get transmitted instantaneously to all our factories. That's really the power of how we attack our yield problems from an organizational standpoint.

Q: So the engineers in your Chandler (AZ) fab have a particular defect issue that's new, they discover what it is, get to the sources and mechanisms, the prevention and elimination aspects of it. As soon as that defect problem is solved, then the rest of the people have that solution in their library?

A: No, they have to implement it.

Q: They have to implement it immediately whether they've seen the problem in their fab or not?

A: Yes.

Q: What kind of differentiations do you see between defect reduction and process control?

A: I think those two are really quite different. Process control is the responsibility of the manufacturing organization, to keep the process where it's supposed to be, monitor it properly, and do the things to keep it under control. Defect reduction is an improvement effort where you focus on making improvements to the manufacturing process.

Q: How have you seen the role and relationship between you and your suppliers change in terms of technology, delivery, sales, and service?

A: I'd say a couple of things. One is that the sophistication of the suppliers has become dramatically higher. The level of technology that they have to deal with has increased dramatically so their technical capability has increased, so in that sense the interaction between the suppliers and us is much richer than it was in the past. And it's not just a commercial negotiation, it's really a technical assessment, a technical negotiation as well as a commercial assessment. That's a big change over the last 10 to 15 years.

Q: Is there anything in the last few years that really jumps out at you as far as the relationship is concerned?

A: I think the collaboration between the semiconductor companies and the equipment companies has changed a lot. We're doing many more joint projects than we ever did in the past. That's where I see a big difference. The whole move for process equipment companies to do their cell or total module development is coming from a different location.

Q: Would you attribute some of the collaboration to the effect Sematech has had on the industry?

A: Sematech has provided an umbrella for the semiconductor and semiconductor equipment companies to work much more closely together. It has allowed the companies to get to know each other on a different type of basis than just a customer-supplier negotiation basis, by working on issues of the industry and of the overall ITRS roadmap.

Q: If you had a personal wish list, what would be the top two or three things you'd like to see your suppliers improve?

A: The three things that are important to us are technology, cost, and agility. Really getting the technology when we need it, and getting it there ahead of where we need it, having the right cost and productivity of the equipment, and then being able to move very quickly with us as we ramp capacity across our big manufacturing network.

Q: Now the suppliers might say, "If we could get your final decisions quicker, then we could project out farther, reduce our cycle times, reduce your lead times...."

A: I don't think that has anything to do with their cycle times. It's intrinsic in how the equipment is designed and made and the supply chain that they have. That's what I am trying to encourage the equipment companies to go look at and redesign, so their cycle time can really be significantly less than it is today.

Q: Aside from lithography, what other parts of the production line do you believe need a more intense effort?

A: Gate dielectric is extremely important. We're down to 20 angstroms gate dielectric. We're going to run out of voltage scaling fairly shortly, so that means that we have to do something with the materials we're using for gate oxides. I think there's a huge concern coming up, so we need some people to go work on that and solve that problem.

Q: What about some of the back-end-of-line copper metallization issues. Do you see those as potential roadmap slowers at all?

A: I don't think so. I think there are enough people and companies working on the copper solutions that it should be ready for us to move quickly into our production in our 0.13-µm generation.

Q: I wanted to ask you about the old Digital fab in Hudson (MA). What were some of the challenges in that fab's transition into the Intel fold?

A: A lot of the things are cultural, so we have done a lot of training and had a number of people go to existing Intel fabs, to see how those operate. We also sent a number of people to the Hudson operation. So that interchange over a period of time really gets the organization to understand how Intel operates its fabs and what we expect in productivity, cost, and technology.

Q: Were there aspects of their original operations that you integrated into your operations?

A: We certainly learned and incorporated a number of technology and equipment details that were quite useful for us in how we operate.

Q: What do you think of the blurring of the traditional front end of wafer processing and the test, assembly, and packaging side that we're starting to see?

A: That's a product and technology specific situation. If you're thinking about packageless products and the like, I think that's going to happen in memories and small-form-factor products first. For products with huge pinouts like chip sets, network processors, and microprocessors, that's well down the road. There are huge material-interface problems to be able to resolve that. On top of that, there's a strong desire for our microprocessor customers to have the parts socketable, so that they can customize their product at the end of their manufacturing.

Q: So it's not on your roadmap as close as it is on some other peoples'?

A: We're always trying to optimize our testing on what we test and sort, and what we test at final test. But I don't see us getting away from that and going to die-only for many of our products. Certainly in memory, it's a possibility, and we're working in that direction.

Q: Is there any particular part of the process technology that you're more versed in from your own initial training?

A: You have to remember that when I was a process engineer, you could do everything, you could actually make an integrated circuit. Now, no one person could even come close to making an integrated circuit.

Q: Is there any one part of the line that especially interests you?

A: Lithography is incredibly fascinating, it has the highest leverage and most importance. Just remembering back to 1980 or so when the first steppers came on the scene, we were wondering whether we could even go below 1 µm with optical stepper technology. To see the changes that we've gone through to get down to 0.18 µm and below with basically the same kind of technology is absolutely incredible.

Q: What's the favorite part of your job?

A: The favorite part of my job is when I see the output of a huge organization, and how a huge organization is able to change so quickly. It gives me a lot of satisfaction to see how fast things change in this industry and see that our organization responds quickly to the needs of our customers.

MicroHome | Search | Current Issue | MicroArchives
Buyers Guide | Media Kit